Model | Resource | Character | Samples | Language | Source | Method |
---|---|---|---|---|---|---|
PDP | [paper], [--] | 327 | 1,042,647 | EN, ZH | TV shows | Experience Extraction, Dialogue Synthesis |
Character-LLM | [paper], [data] | 9 | 1,900,800 | EN | Encyclopedia | Experience Extraction, Dialogue Synthesis |
ChatHaruhi | [paper], [data] | 32 | 54,726 | EN, ZH | Books, Games, Movies | Experience Extraction, Dialogue Synthesis |
RoleLLM | [paper], [data] | 100 | 140,726 | EN, ZH | Scripts | Experience Extraction, Dialogue Synthesis |
HPD | [paper], [data] | - | 1,191* | EN, ZH | Books | Dialogue Synthesis, Human Annotation |
CharacterGLM | [paper], [data] | 250 | 1034* | ZH | Books, Scripts | Experience Extraction, Dialogue Synthesis, Human Annotation |
PIPPA | [[paper],[data] | 1,254 | 25,940* | EN | Character.ai-Users | Dialogue Synthesis |
RoleEval | [paper], [data] | 300 | 6,000 | EN, ZH | Encyclopedia | Dialogue Synthesis |
CharacterEval | [paper], [data] | 77 | 11,376 | ZH | Books, Scripts | Experience Extraction, Human Annotation |
CroSS-MR | [paper], [data] | 126 | 445 | EN | Literature-Analysis | Experience Extraction |
DITTO | [paper], [data] | 4,002 | 36,662 | EN, ZH | Encyclopedia | Experience Extraction, Dialogue Synthesis |
RoleInteract | [paper], [data] | 500 | 30,800 | EN, ZH | Books, Movies | Experience Extraction, Dialogue Synthesis |
LifeChoice | [paper], [data] | 1,401 | 1,401 | EN | Literature-Analysis | Experience Extraction |
InCharacter | paper], [data] | 32 | 18,304 | EN, ZH | Personality-Tests | Dialogue Synthesis |
[1] INCHARACTER: Evaluating Personality Fidelity in Role-Playing Agents through Psychological Interviews. [paper], [code]
[2] Character-LLM: A Trainable Agent for Role-Playing. [paper], [code]
[3] RoleLLM: Benchmarking, Eliciting, and Enhancing Role-Playing Abilities of Large Language Models. [paper], [code]
(1) Raw groundtruths of general instructions without role-playing (RAW); (2) Customized general instruction responses with role-playing from RoleBench-general (CUS); (3) Role-specific instruction responses from RoleBench-specific (SPE).
[4] ChatHaruhi: Reviving Anime Character in Reality via Large Language Model. [paper], [code]
[5] CharacterEval: A Chinese Benchmark for Role-Playing Conversational Agent Evaluation. [paper], [code]
[6] Toxicity Evaluation Toxicity in CHATGPT: Analyzing Persona-assigned Language Models. [paper]
[7] NarrativePlay: Interactive Narrative Understanding. [paper]
Please cite the following papers as the references if you use our codes or the processed datasets.
@article{xxxxx,
title={xxxxxx},
author={xxx},
journal={arXiv preprint arXiv:xxx},
year={2024}
}